ESL Pathways: Mapping the Whole Student - CAIR · 2020-03-18 · ESL Pathways: Mapping the Whole...
Transcript of ESL Pathways: Mapping the Whole Student - CAIR · 2020-03-18 · ESL Pathways: Mapping the Whole...
ESL Pathways: Mapping
the Whole Student
California Association for Institutional Research
November 7, 2019
Craig Hayward, Bakersfield College + The RP Group
Terra Morris, The RP Group
Mallory Newell, De Anza College + The RP Group
Terrence Willett, Cabrillo College + The RP Group
Overview
• Sankey Diagrams: Mapping ELL Student Pathways
• AB 705 Implications for ESL Students
• English Language Learner (ELL) Types
– International Students
– ELL US HS graduates
– Other Degree/Transfer-seeking ESL students
• Discussion
• Visualization technique used to display flow or movement
• Allows you to show complex processes visually
• Width of lines are proportional to overall quantity represented in
the diagram
• Interactive
• Powerful conversation starters
Why Sankey Diagrams?
• Data source: CCC Chancellor’s Office MIS data, credit and noncredit
course enrollments between fall 2012 and winter 2019
• Used Microsoft SQL Server Management Studio 17 to modify, quantify, and
create a final data file matching required specifications for Sankey diagram,
tracking CC ESL students’ pathways and progress through the ESL and
English sequences across six academic terms
• Researched and tested multiple free and trial versions of visualization
software capable of creating Sankey diagrams (R, Displayr, floWeaver)
• Selected Google Charts Sankey package: free, simple and straightforward,
easy to modify html code to add titles, colors and features using Notepad++
Method
R Google js(abridged)
<script type="text/javascript" src="https://www.gstatic.com/charts/loader.js"></script>
<div id="sankey_multiple" style="width: 10000px; height: 100%;"></div>
<script type="text/javascript">
google.charts.load("current", {packages:["sankey"]});
google.charts.setOnLoadCallback(drawChart);
function drawChart() {
var data = new google.visualization.DataTable();
data.addColumn('string', 'From');
data.addColumn('string', 'To');
data.addColumn('number', 'Weight');
data.addRows([
[ "CR ESL one level below", "Term1 TLEE", 581 ],
[ "CR ESL one level below", "Term1 CR one level below", 10853 ],
[ "Term1 TLEE", "Term1 TLEE success", 458 ],
[ "Term1 TLEE", "Term1 TLEE non-success", 123 ],
[ "Term1 CR one level below", "Term1 CR1 success", 8224 ],
Sample Google Charts html script used to create Sankey
diagrams
1. English Language Learner (ELL) U.S. High School
Graduates
2. International Students (IS)
3. Non-IS, non-U.S. high school graduate ESL students who
are degree/transfer seeking
Degree/Transfer-seeking ESL Student Types
that are Affected by AB 705
37%of ESL students fall
into one of these
three student types
For more on AB 705 go to:
https://assessment.cccco.edu/ab-705-implementation
ELL/ESL Throughput Data File Description
• 92 cohorts of students tracked forward for three years or,
sometimes, for one year (throughput cohorts)
– 555,625 noncredit students
– 328,403 credit students
• Focal sample time period 2004 through 2018
• Further disaggregated by student types and/or
degree/transfer-seeking status
U.S. High School Graduates
High School English Language Learners
• How do ELL students with less than 4 years of high school fare in
transfer-level English?
• Compare ELL high school graduates who started at one-level
below in ESL to those who started in transfer-level English.
• Citizenship status is not related to US high school students’
throughput in transfer-level English.
U.S. High School Graduates’ Transfer-level English Throughput
by Years in U.S. High School, ELL Designation, and Community
College English Language Arts Pathway (excludes US Citizens)
30% 30%25% 26%
31% 33% 29%35%
45% 48% 49% 48%
81% 83% 81% 82%
0%
20%
40%
60%
80%
100%
Just Sr. Jr.+Sr. So.+Jr.+Sr. Four years
ELL in HS, ESL in CC Not ELL designated in HS, ESL in CC
ELL in HS, English in College 1 yr. post-AB 705 TLC throughput (projected)
ELL US HS Graduates Have Higher Throughput on
the English Path than the ESL Path
86%
18%
87%
21%
85%
20%
84%
19%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Senior year only(English path)
Senior year only(ESL path)
Jr. + Sr. (Englishpath)
Jr. + Sr. (ESL path) So., Jr. & Sr.(English path)
So., Jr. & Sr. (ESLpath)
All 4 yrs. (Englishpath)
All 4 yrs.
One-year TLC Throughput for ELL US High School Graduates Disaggregated by English vs. ESL Path, All Citizenship Statuses, and Years of U.S. High School
US Citizen Permanent Resident Temporary Resident Refugee/Asylee Student Visa (F-1 or M-1) Other Unknown Total
High School English Language Learners• Can we account for differences in high school GPA between
students who started in transfer-level English composition at the
community college and those who did not?
• Derive expected throughput of ELL high school graduates who
start in transfer-level English after adjusting for differences in
high school GPA.
• Derive expected one-year throughput rate of students who start
in one-level below in ESL and then transition to transfer-level
English composition.
20%25%
38%
84%
69%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
ESL path (n =2,296)
Only at collegesallowing directtransition into
TLE (n = 1,725)
Second classactually was
TLE (n = 708)
English path -Starting at TLE
(n = 33,491)
English path -Starting at TLE(GPA weighted
estimate)
1-yr. Throughput of ELL US HS Graduates: 1-level below ESL vs. Transfer-level English
ESL path from 1 level below - One-yr. throughput rate
Transfer-level English path - One-yr. throughput rate
English Language Learners:
ESL vs. English Pathways
Discussion• Given how common it is for ELL US HS grads to enter
mainstream English, what types of ESL support will
improve the success of English Language Learners?
• What does the evidence suggest regarding which English
Language Arts pathway will maximize the probability of
completion of transfer-level English composition?
• Is the ESL label particularly problematic for generation
1.5 students, as suggested by Ortmeier-Hooper (2008)?
International Students
International Student TLE Throughput Rates
ESL path = 28,584
English path = 16,958
22%
34%
50%
58% 61%
95%
29%
41%
62%66%
88%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
5+ levels below 4 levels below 3 levels below 2 levels below 1 level below Transfer level
ESL path English path
Majority of International Students on English
Path Start at Transfer-level
ESL path = 28,584
English path = 16,958
5%
13%
28% 27%25%
1%0.5%3%
19%
25%
53%
0%
10%
20%
30%
40%
50%
60%
5+ levels below 4 levels below 3 levels below 2 levels below 1 level below Transfer level
ESL path English path
Asian Language Group Represents Large
Majority of International Students
ESL path = 28,584
English path = 16,958
Asian, 58%
MENA/ European, 12%
Hispanic, 3%
African, 2%
East Indian, 1%
Filipino, 1%
Other, 1%
Unknown, 23%
Other Degree/Transfer-seeking ESL StudentsNon-International, Non-US HS Graduates
Third Group ESL Students
• Do some colleges have higher throughput rates for ESL
students who do not have US high school diplomas and
who are not international students?
• If so, what is driving the difference in throughput rates?
College’s Average ESL Starting Level and
Throughput Rate are Highly Correlated
R² = 0.504
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
-6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00
Avera
ge T
LE
Thro
ughput
in 3
yrs
.
Average ESL Levels Below
N = 47 colleges with 30+ students of
this type
Credit ESL students with no diploma
Third Group ESL Students
• To what extent are demographic differences among the
groups of degree/transfer-seeking “third group” ESL
students related to the observed differences among
colleges in throughput rates?
College throughput rates for degree-seeking
ESL students with no diploma
0%
10%
20%
30%
40%
50%
60%
-10% 0% 10% 20% 30% 40% 50% 60% 70%
TL
E th
rou
gh
pu
t fr
om
ES
L (
co
lleg
e m
ea
n)
Regression Adjusted (Press) Predicted Value
N = 47 colleges with 30+ students of this type – no U.S. HS diploma
Adj. R2 = 0.68
Independent Variable = Starting ESL level
Controls = age, language group, citizenship status
What lessons can we draw?
• There is an interplay between ESL placement practices
and ESL curriculum such that at colleges where third
group ESL students are typically placed into higher
levels, the average throughput rate is commensurately
higher.
Inter-college variation in average starting ESL level and TLE
throughput for degree/transfer-seeking students with no US diploma
N = 88 colleges with 30+
“third group” students
16%
25%
31%
0%
5%
10%
15%
20%
25%
30%
35%
< 4 levels below >= 4 and < 3 levelsbelow
>= 3 levels below
Three-year Throughput Rates
Third Group Analysis Highlights
• Not all colleges have a significant “third group” population
• Some colleges are more effectively moving this group of
students to and through TLC than others
– How do assessment, placement and curricular structures
correlate with higher levels of student throughput?
• Colleges that place most “third group” students at or around 3 levels
below have the highest average throughput for this group (27%)
• College throughput rates drop by about ten percentage points for
each level below three that colleges place most students (i.e., 17%
for 4 levels below, 7% for 5+ levels below)
Discussion
MMAP Research Team
Terrence Willett
The RP Group
Mallory Newell
The RP Group
Craig Hayward
The RP Group
Inge Bond
The RP Group
John Hetts
CCCCO
Daniel Lamoree
ERP
Andrew Kretz
The RP Group
Loris Fagioli
The RP Group
Addendum: Descriptive slides
Relative sizes of the various ESL student subtypes
Relative Sizes of First-time English Language Arts Pathways in 2017-18 for
Degree/Transfer Seeking ELL/ESL Students
About 87% of ELL US High
School graduates enter
mainstream English at the
community college.
TLC Throughput by Credit/Noncredit and
Student Journey Type
35%
11%8%
12%
1% 1%0%
5%
10%
15%
20%
25%
30%
35%
40%
Degree-Transfer Short term CTE Adult Ed
Credit Noncredit
Nota Bene: AB 705 only
applies to ESL students
who are degree/transfer-
seeking.